Large-Scale Chemical and Crystal Packing Space Exploration for HOF-Based Porous Materials

Functional materials impact most aspects of our lives, but their design at the microscopic scale cannot be approached with the engineering rules that would be used in planning and constructing macroscopic objects. The ADAM (Autonomous Discovery of Advanced Materials) project aims at changing the way to discover functional molecular crystals by coupling a computational engine for the evolutionary exploration of chemical space using Crystal Structure Prediction (CSP) and Machine Learning, with an experimental engine for autonomous synthesis and characterization using robo-chemists. Although the ADAM project has a much broader scope, we here focus on a computational screening strategy for porous molecular materials based on hydrogen bonding groups from known Hydrogen-Bonded Organic Frameworks (HOFs). Porous packing is a property that cannot be predicted purely based on molecular structure, so requires CSP on all candidate molecules. This contribution will discuss how this is now possible for candidate libraries of thousands of molecules.